Environmental Biology of Fishes 66: 391–400, 2003.
© 2003 Kluwer Academic Publishers. Printed in the Netherlands.
Effects of barriers and thermal refugia on local movement of
the threatened leopard darter, Percina pantherina
Jacob F. Schaefera,e , Edie Marsh-Matthewsa,b , Daniel E. Spoonerc , Keith B. Gidoa,f & William J. Matthewsa,d
Sam Noble Oklahoma Museum of Natural History, University of Oklahoma, Norman, OK 73072, U.S.A.
b
Department of Zoology, University of Oklahoma, Norman, OK 73072, U.S.A.
c
Oklahoma Biological Survey, University of Oklahoma, Norman, OK 73019, U.S.A.
d
Biological Station, University of Oklahoma, Kingston, OK 73439, U.S.A.
e
Present address: Department of Biology, Southern Illinois University Edwardsville, Edwardsville,
IL 62026, U.S.A. (e-mail: jacscha@siue.edu)
f
Present address: Division of Biology, Ackert Hall, Kansas State University, Manhattan, KS 66506, U.S.A.
a
Received 28 June 2001
Accepted 15 August 2002
Key words: habitat, conservation, culverts, percids, fish
Synopsis
Local, short-term dispersal by the U.S. federally-threatened leopard darter, Percina pantherina, was examined in
the field and in the laboratory to assess the possible effects of natural versus man-made barriers on movement.
Mark-resight studies were conducted in two summers at sites in the Glover River (southeastern Oklahoma, U.S.A.).
At one site, patches of ‘preferred’ habitat were separated by a natural riffle; at the other site, by a low-water road
crossing with culverts. At the Natural Riffle site, darters moved downstream across the riffle, but also moved upstream
into deeper water when water temperatures exceeded 29◦ C in the ‘preferred’ habitat. Use of deeper, cooler waters by
this species in late summer suggests that thermal refugia may be important habitats for the long-term management
of leopard darters. At the Road Crossing site, all documented movement was in a downstream direction, and at
least two darters traversed culverts in the low-water bridge. Laboratory studies of movement across several types
of culverts suggested that culverts significantly decrease the probability of movement among habitat patches.
Introduction
Management to protect threatened and endangered
stream fishes requires preservation of preferred or optimal habitat, the full spectrum of habitat types used
and the corridors that allow movement among them
(Schlosser 1995, Smithson & Johnston 1999). The
importance of long-distance movement for spawning (e.g., Colorado pikeminnow, Ptychocheilus lucius,
Tyus 1990) or seasonal avoidance of habitat types
of diminished quality (e.g., brown trout, Salmo
trutta, Clapp et al. 1990) is documented for many
stream fishes. The significance of short-term, localized
movements among habitats is becoming recognized
increasingly because such movements are important for habitat assessment (Power 1984), foraging
(Clapp et al. 1990), escape from predators (Power et al.
1985, Harvey 1991), or use of thermal refugia (Kaya
et al. 1977, Matthews & Berg 1997). Studies of movement in many species of stream fishes reveal that typically only a few individuals in a population move
long distances (Smithson & Johnston 1999, Schaefer
2001) while most individuals tend to remain in a home
pool but make exploratory forays into adjacent habitat (Smithson & Johnston 1999). These short-term,
localized movements among habitats suggest that even
sedentary stream fishes may require larger areas of
habitat than typically assumed (Smithson & Johnston
1999).
This study examined effects of road crossings on local movements of the U.S. federallythreatened leopard darter, Percina pantherina, to make
392
recommendations about habitat management for this
species in the Glover River, Oklahoma. Percina
pantherina was listed as threatened by the U.S. Fish
and Wildlife Service in 1978, and critical habitat was
designated in three major tributaries of the Little River
system in Oklahoma and Arkansas (U.S. Fish and
Wildlife Service 1978). The historic range of this
species encompasses five of the six major drainages of
the Little River system in southeastern Oklahoma and
southwestern Arkansas (Miller & Robison 1973, Eley
et al. 1975, Zale et al. 1994). Four of the five known
populations are restricted to upland reaches and are isolated from other populations by downstream reservoirs
(Zale et al. 1994, Williams et al. 1999). The only unimpounded drainage with the leopard darter is the Glover
River, McCurtain County, Oklahoma (Zale et al. 1994,
Williams et al. 1999). The population in this drainage is
relatively large (Zale et al. 1994, Williams et al. 1999),
although recent surveys by U.S. Fish and Wildlife
Service and United States Department of Agriculture
(USDA) Forest Service suggest that the Glover River
population is declining (K. Collins unpublished data).
Much of the critical habitat in the Glover River currently is managed by the USDA Forest Service with
goals that include maintaining thriving populations of
leopard darters. To this end, the Forest Service sought
information to determine the need for modifying or
removing road crossings that might inhibit leopard
darter passage among habitat patches (R. Standage,
personal communication).
Many low-water crossings allow water to pass
through round pipe or square-to-rectangular box culverts. These culverts vary in size, but all potentially
reduce migration of fishes by concentrating discharge
and reducing the navigable cross-sectional area fish
use to move within the stream. Pipe culverts in particular significantly reduce movement of stream fishes,
although other types of openings may not impact movement (Warren & Pardew 1998). Laboratory studies of
swimming ability suggest that current velocity in both
pipe and box culverts restricts movement of leopard
darters (Toepfer et al. 1999), although rates of movement under natural conditions are unknown for this
species. The degree to which leopard darters move
across barriers, either natural or artificial, is likely
to depend on the characteristics of the barrier itself.
Current velocity, riffle length, and/or thalweg depth
affect movement of various stream fish across natural
riffles (Lonzarich et al. 2000, Schaefer 1999, 2001).
Furthermore, different species show differential rates
of movement across both natural and artificial barriers
(Warren & Pardew 1998, Schaefer 1999, Lonzarich
et al. 2000).
We conducted a set of field mark-resight surveys
and laboratory experiments designed to measure the
effects of road crossings on leopard darter movement.
In field studies, we compared movement rates across
road crossings and over natural riffles in the Glover
River and documented movement into deep-water habitats in response to water temperature variation. In laboratory experiments we used scaled-down culverts in
a large artificial stream to assess effects of different
culvert characteristics on leopard darter movement.
Methods
Field survey and resight methods
We selected two sites on the Glover River known to
have large leopard darter populations based on surveys by the United States Fish and Wildlife Service
(K. Collins unpublished data). Federal permit restrictions limited the ‘take’ of leopard darters during the
2 years of the field survey to 14 individuals. Although
we considered surveying four sites (two each with and
without culverts), we chose to restrict the number of
sites and mark more darters, and thereby maximize
sample size, at each site.
At both sites, the river channel was narrow with
distinct riffle and pool structure. Substrate consisted
mostly of large cobble, boulders and some reaches with
bedrock. The reported ‘preferred’ or ‘optimal’ habitat for the leopard darter is large cobble to boulder
substrate in low or non-flowing water less than 1 m
deep (Jones et al. 1984, James et al. 1991), and it has
been generally thought that the species is restricted to
these habitats, except in spring when they move into
gravel riffles to spawn (Jones et al. 1984, James &
Maughan 1989, James 1996). Based on these published descriptions of preferred habitat, we concentrated survey and marking effort around the edges of
pools where there were boulders and some emergent
vegetation. One site (hereafter called Natural Riffle)
had no man-made road crossing, while the second site
(Road Crossing) had a low-water road crossing with
two round culverts roughly 60 cm in diameter and four
box culverts approximately 3 m wide (Table 1).
At each site, we identified four patches (numbered
1–4 from upstream to downstream). Patches 2
and 3 were areas known to be suitable habitat for
leopard darters based on previous surveys and habitat
393
Table 1. Habitat characteristics for the Natural Riffle (NR) and Road Crossing (RC) sites on the Glover River,
McCurtain County, Oklahoma. At each site, four patches were identified and numbered 1–4 from upstream to
downstream. For the mark-resight study, leopard darters were captured from patches 2 and 3, marked and released
into the patch of origin. Resight surveys were conducted in all four patches. Depths given in the description indicate
relative depths of the patches because absolute depth varied over the course of the study.
Size (length × width m)
Site
Patch
NR
1
15 × 30
Habitat between patches = 20 m span of bedrock habitat
2
80 × 25
Natural Riffle Barrier = 20 m natural riffle
3
35 × 20
Habitat between patches = 10 m natural riffle
4
100 × 20
Deep pool, boulder
1
25 × 30
Habitat between patches = 3 m natural riffle
2
85 × 20
Road Crossing Barrier = road crossing with culverts
3
20 × 25
Habitat between patches = 15 m natural riffle
4
100 × 15
Shallow, boulder, cobble
NR
NR
NR
RC
RC
RC
RC
characteristics (Table 1). At the Natural Riffle site, a
riffle approximately 20 m long separated patch 2 from
patch 3. At the Road Crossing site, the road separated
patch 2 from patch 3. At each site, we captured darters
for marking in patches 2 and 3. For marking, snorkelers hand-netted fish and transported them to 40 liter
containers along the shore. Using a standard 27 gauge
syringe, we gave fish a small subcutaneous injection
of acrylic paint (Liquitex brand: red, orange, yellow
and blue) following the method of Hill & Grossman
(1987). To allow us to assess movement across barriers, we used different colors to mark fish from patches 2
and 3. Other studies (e.g., Schaefer 1999) demonstrated
that these marks are visible for up to 2 months and that
stress to fish is minimal. Only six mortalities of 266
marked fish were directly attributable to the marking
procedure. After fish recovered, we released them back
into the patch of origin. We released all fish from a
given patch at the same point near the middle of that
patch.
In the weeks following marking, we conducted
snorkel and SCUBA surveys to look for marked fish.
In each patch, observers swam non-overlapping transects across the stream (starting downstream and working upstream) covering the entire area. We surveyed
patches 2 and 3 for 2 person-hours each and patches 1
and 4 for 1 person-hour each. Matthews et al. (1994)
found that a survey time of 15–20 min by a single
snorkeling observer was sufficient to detect most fishes
in large pools of a south Oklahoma stream. We counted
all fish, marked and unmarked, in a given patch.
Description
Shallow, boulder, cobble, gravel
Medium depth, cobble, gravel, silt
Medium depth, cobble, gravel
Medium depth, boulder, cobble, gravel
Shallow, deep in areas, boulder, cobble
Shallow, boulder, cobble
In 1999 (4–5 August), we marked and released
79 individuals at the Natural Riffle site (45 from
patch 2 and 34 from patch 3) and 35 individuals at
the Road Crossing site (22 from patch 2 and 13 from
patch 3). We conducted surveys for marked individuals on 15 August, 28 August, 18 September, 8 October,
and 21 October 1999. In 2000 (30 June to 1 July), we
marked and released 87 darters at the Natural Riffle
site (57 from patch 2 and 30 from patch 3), and 59 at
the Road Crossing site (32 from patch 2 and 27 from
patch 3) (Figure 2). We conducted resight surveys on
7 July, 12 July, and 20 July 2000.
Laboratory methods
On 8 October 1999, we hand-netted 20 leopard darters
(10 from each field site) and transported them to the
University of Oklahoma Research Park in Norman,
OK, for laboratory trials. We marked individual fish
uniquely with acrylic paint injections. We maintained
fish from October 1999 until August 2000 in holding
pools 1.3 m in diameter and 15–20 cm deep, and fed
them daily with a mixed diet of frozen brine shrimp
and frozen bloodworms.
Our indoor artificial stream was 8 m long, 1.3 m
wide, and averaged 40 cm water depth. The stream
was made as large as possible so that habitat patches
established within could be large and widely spaced.
There was only room for one such stream in the available laboratory space. The stream substrate consisted
of gravel, cobble, and boulders (13–26 cm in diameter)
394
from the Natural Riffle site on the Glover River. Three
equally-spaced patches (hereafter designated Upper,
Middle, Lower), designed to imitate the preferred
habitat of Percina pantherina were established in the
stream. Each patch consisted of 5 boulders in an array
roughly 1.5 m in diameter. The Upper and Middle
patches were separated by a barrier constructed of
cement blocks with a middle section (culvert) that
could be altered according to treatment. Although the
barrier could be reconfigured to test effects of culvert
shape on movement, the basic structure of the barrier
was permanent so the position of the barrier could
not be changed between trials. There was no barrier
between the Middle and Lower patches. Water was
pumped from one end of the stream to the other by
one to three sump pumps (Little Giant Pump Company,
Oklahoma City, model 6-CIM-R) each rated at
10 410 liter h−1 .
Treatments consisted of four different culverts
(Table 2): square-narrow, square-wide, round-smooth,
and round-ribbed. Culverts used in laboratory trials
were smaller than those at the Road Crossing site, but
they presented effects on stream flow similar to those in
the Glover River. Schwalme et al. (1985) and Katapodis
et al. (1991) found that even large fish (>500 mm FL)
pass though Denil fishways only 30 cm wide (less than
half the width of our square-wide laboratory culvert).
While these culverts were smaller than many in natural
streams, the thalweg depths and current velocities in
our trials were similar to those measured in the field.
In addition, a darter in one of our culverts took up
much less than 25% of the total cross-sectional area of
the culvert so that flows were not restricted. Typically,
objects taking up 25% or more of this cross-sectional
area will begin to restrict flow and cause compression
forces (Suren et al. 2000).
For each culvert type, we conducted one trial with
1 pump in operation (low discharge) and a second trial
with 3 pumps in operation (high discharge). For all
culvert types except the square-wide, water velocity
inside the culvert was 15 cm s−1 for the low discharge
trials and 30 cm s−1 for the high discharge trials. For the
square-wide culvert, which had a much larger crosssectional area (Table 2), water velocity was 7.5 cm s−1
for the low discharge trial and 15 cm s−1 for the high
discharge trial.
We conducted one 10-day trial for each dischargeculvert shape combination (eight total trials). Because
we were limited by permit restrictions to holding only
20 leopard darters in the laboratory, individual darters
were used in more than one trial. For each trial, we
selected nine individuals haphazardly from the holding tank and released three fish into each of the three
patches of the artificial stream. For each 10-day trial,
we made daily observations and recorded the location (patch) of each fish observed. From these observations we calculated the minimum number of moves
necessary to explain changes in the daily distribution
of darters in the artificial stream. We classified movements as upstream or downstream, and as crossing the
barrier or not. During the trials, we fed fish (about
noon) by distributing equal amounts of food to each
patch. We carried out surveys in the evening of the
10 days following the initiation of the trial. On average, we saw 60% of marked fish in any one survey,
almost always within the boulder habitat patches and
not in the space between patches. After each trial,
we removed all fish and returned them to holding
pools.
Results
Field surveys and resights
Table 2. Conditions of the passage through the barrier for each
of the four trials. The barrier was located between the upper and
middle patches and was constructed of cement blocks. Thalweg
depth remained constant at 5 cm.
Trial
Shape
Width
(cm)
Length
(cm)
Current
(cm s−1 )
1
2
3
4
5
6
7
8
Square-narrow
Square-narrow
Square-wide
Square-wide
Round-smooth
Round-smooth
Round-ribbed
Round-ribbed
19
19
76
76
16
16
16
16
38
38
38
38
38
38
38
38
15
30
7.5
15
15
30
15
30
During the eight resighting surveys, we counted a total
of 1251 leopard darters. At both study sites, the number of leopard darters observed and the distribution of
fish among patches varied over time (Figure 1). At
the Natural Riffle site (Figure 1a–d), we most often
observed leopard darters in patch 2 (an area considered
to be ‘optimal’ habitat based on depth, substrate, and
current), but we also found darters in large numbers
in patches 1 and 3. Patch 3 was similar to patch 2,
but patch 1 was much deeper (4–5 m deep) than is
considered ‘optimal’ for leopard darters. We recorded
extensive use of patch 1 at times (15 August 1999,
395
Figure 1. Total number of leopard darters seen during snorkel and SCUBA surveys in 1999 and 2000, in each patch (numbered 1–4,
upstream to downstream) at the Natural Riffle site (a–d) and the Road Crossing site (e–h).
28 August 1999, and 20 July 2000) when water temperature in the shallower patches was very warm
(approximately 30◦ C or higher) and temperatures in the
deeper pool were 3–4◦ C lower. At the Road Crossing
site, we observed most leopard darters in patch 2 (again,
an area of ‘optimal’ habitat), although darters were also
observed in patches 3 and 4 (Figure 1e–h). At this site,
there were no available deep-water habitats in any of
the four patches.
In both years combined, we marked a total of 260
leopard darters. In the eight surveys after marking, we
resighted marked fish 97 times. Most marked darters
were resighted in the patch where they were originally
marked and released (Figure 2), but 11 resights were
in patches other than the patch of origin.
At the Natural Riffle site (Figure 2a–d), we recorded
movement downstream (but not upstream) across the
natural barrier. At least two (and possibly three, if all
396
Figure 2. Number of leopard darters marked (designated by ‘M’ above the date of marking) and resighted in 1999 and 2000 at the Natural
Riffle site (a–d) and the Road Crossing site (e–h). Darters marked in patch 2 (upstream of the barrier) indicated by black bars; darters
marked in patch 3 (downstream of the barrier) indicated by open bars.
sightings on different dates represented different individuals) darters marked in patch 2 were resighted
downstream below the natural riffle in patch 3 in 1999.
Darters did move upstream into the deep pool habitat.
In 1999, at least two (and possibly as many as five)
darters moved upstream from patch 2 to patch 1. In
2000, at least three darters (and possibly four) moved
from patch 2 to patch 1. Movement from patch 2 to
patch 1 required moving upstream across a 20–30 m
stretch of bedrock substrate.
At the Road Crossing site (Figure 2e–h), all documented movement was in a downstream direction. In
2000, at least one (and possibly two) darters crossed
the road crossing barrier as well as a natural riffle from
397
patch 2 to patch 4, a distance of at least 200 m. One
other marked darter moved from patch 3 to patch 4.
Most movement among patches was from patch 2 to
patch 1 at the Natural Riffle site. We resighted marked
darters that had moved upstream into this deeper habitat at times when large numbers of unmarked leopard
darters also occupied this habitat (Figures 1 and 2).
Occupancy of the deeper habitat coincided with higher
temperatures (32–34◦ C) in the surrounding shallow
habitat. Temperature profiles of the deep hole and surrounding shallow habitats at the Natural Riffle site
were measured on three dates in 2000: 1 July (the
day that fish were marked), 7 July, and 20 July. On
1 July, temperatures of the shallow habitats (10–80 cm
deep) ranged from 25.5◦ C to 27◦ C. Temperature of the
deep hole ranged from 26◦ C at the surface to 23◦ C
at 5–6 m deep. At this time, no leopard darters were
observed in the deep hole, which was occupied by a
large bass (Micropterus sp.). On 7 July, water temperature at shallow (<10 cm deep) areas was 32◦ C, but
temperature in typical leopard darter habitat (ca. 80 cm
deep) was 29◦ C. Temperature in the deep pool was 26–
27◦ C. Four leopard darters were observed in this deep
habitat on this date. On 20 July, water temperatures
were much warmer than in the previous surveys. Water
temperature decreased with depth in patch 1 as follows: 33◦ C at 0.3 m, 30◦ C at 1.2 m, 29◦ C at 3 m, 28◦ C
at 3.3 m. Most leopard darters in patch 1 at this time
were observed at a depth of 3 m.
Figure 3. Number of moves per trial (±1 SE) for all trials (n = 8)
in the artificial stream in an a-upstream versus downstream direction and b-across a barrier versus between patches without a
barrier.
Laboratory studies
Darters in all trials actively dispersed among all three
patches of the artificial stream. Using combined data
from all trials, movement upstream (mean = 11.1
moves per trial, SE = 1.9) did not differ from that
downstream (mean = 9.6 moves per trial, SE = 0.80,
Figure 3a) (paired t = 1.15, p = 0.26, df = 7, power =
0.19). The barrier (regardless of culvert type), however, significantly reduced movement (paired t = 6.25,
p < 0.001, df = 7, power = 0.98). Movement between
the two patches with no barrier (mean = 15.9 moves
per trial, SE = 1.9) was much greater than movement
between the patches divided by the barrier (mean = 4.8
moves per trial, SE = 1.2).
Discharge had minimal effect on movement detected
in laboratory trials (Figure 4). Difference in number
of upstream moves at the two discharge levels was
marginally significant with number of moves per trial
at high discharge (mean = 14.25, SE = 2.32) greater
Figure 4. Number of moves per trial (±1 SE) under high discharge (open bars; n = 4 trials) versus low discharge (black bars;
n = 4 trials) conditions for each of the following types of movement: upstream, downstream, across the barrier, and between
barrier-free patches.
than that at low discharge (mean = 8.00, SE = 2.12)
(paired t test, t = 2.7, p = 0.07, df = 3, power = 0.41).
Discharge had no effect on downstream moves (paired
t = 1.5, p = 0.23, df = 3, power = 0.36), movement across the barrier (paired t = 1.5, p = 0.23,
df = 3, power = 0.35), or movement between patches
with no barrier (paired t = 1.5, p = 0.24, df = 3,
power = 0.42).
Movement rates differed marginally across the four
different types of culverts (one-way ANOVA, F =
5.30, p = 0.058, power = 0.44, black bars, Figure 5).
398
Figure 5. Number of moves per trial (±1 SE) in the artificial
stream trials (n = 2, one at high and one at low discharge) for each
of the four culvert types. Black bars represent movement across
the barrier, open bars represent movement between patches with
no barrier.
Movement rate across the square-wide culvert (mean =
9.5 moves per trial, n = 2 trials, one at ‘high’ and one
at ‘low’ discharge, see Table 2) was nearly seven times
that across the round-smooth culvert (mean = 1.5
moves per trial). Movement across the other two culvert types (square-narrow, mean = 3.5 moves per
trial) and round-ribbed (mean = 4.5 moves per trial)
was intermediate. Although trials with different barrier types revealed differences in movement across
the barrier, there were no differences in movement
between patches without a barrier (F = 0.62, p = 0.64,
power = 0.10, Figure 5). This suggests that differences in barrier crossings were, in fact due to barrier
type rather than to variation in movement tendencies
among trials.
Discussion
Leopard darters in the Glover River exhibited limited
movement among habitat patches. Most darters marked
in our field studies in both years were resighted in their
patch of origin, but we did document movements of
up to 200 m. Our observation concurs with those for
other species which suggest that while a small portion
of the population may move long distances, many individuals remain in the same area over the short term
(Funk 1955, Gerking 1959, Freeman 1995; but see
Hill & Grossman 1987, Gowan & Fausch 1996). Given
that our surveys were conducted over a limited area
at each site, we were unable to address long-distance
movements in this study. Of the limited movement
observed in the field, we documented movement in both
upstream and downstream directions, with most being
upstream into deeper habitat when water temperatures
exceeded 29◦ C.
Our field study compared movement across a natural
versus a man-made barrier in a small river. Downstream
movement across the pre-defined barriers (Natural
Riffle and Road Crossing) occurred at both sites but
was limited: only two (or three) individuals crossed the
natural riffle, and only one (or two) individuals crossed
the road barrier (and a natural riffle downstream of
the road crossing). Although upstream movement was
documented for darters moving into cooler waters, no
upstream movement was documented across either of
the pre-defined barriers (natural riffle or culvert) in this
study.
Our laboratory observations suggest that leopard
darters should be able to traverse some types of barriers more readily than others. In artificial stream trials,
leopard darters were more likely to cross square-wide
than narrow (either square or round) culverts. In our
laboratory trials, lowest current speeds across any barrier type were recorded across the square-wide culvert
for both the low and high discharge trials. In the field, it
is also likely that wider culverts increase velocities less
than narrow culverts, and present a larger corridor for
fish passage. During periods of elevated flow, some of
the riffles and culverts in the Glover River had higher
current velocities than we were able to reproduce in the
laboratory. Toepfer et al. (1999) found that both box and
pipe culverts in the Glover River typically had current
velocities greater than 25 cm s−1 (the current velocity at
which they found the greatest swimming activity in the
laboratory). Because the leopard darter exhibits limited
swimming ability, culverts in the Glover River likely
restrict passage under even moderate flow (Toepfer
et al. 1999).
Thermal refugia
Although study of habitat use was not an initial goal of
this study, the serendipitous observation of habitat shift
into deeper, cooler habitats when water temperatures
are high may provide critical insight for conservation
of this threatened species. Our observations suggest
that leopard darters use deep-water habitats as thermal
refugia when temperature in the surrounding shallow
habitats exceeds 29◦ C. The use of thermal refugia by
warmwater stream fishes (Peterson & Rabeni 1996)
and specifically by darters (Smith & Fausch 1997) is a
399
well-studied phenomenon. During our study, flows in
the Glover River (USGS data, site U07337900) varied
from peaks near 283 m3 s−1 to extended periods at or
near 0 m3 s−1 . During these low-water periods, temperatures commonly rose above 34◦ C in habitats where
leopard darters had been very common. At these times,
darters were most abundant in the deep pool habitats
where temperatures were as much as 6◦ C cooler. Daily
fluctuations in stream temperature are extreme in the
Little River drainage. We recorded increases from 28◦ C
to 34◦ C at one location in a matter of hours. This is
especially true for shallow, exposed habitats that are
commonly inhabited by the leopard darter. We often
measured temperatures in shallow areas (36–38◦ C) that
were at or above the thermal tolerance of other darter
species (Smith & Fausch 1997). Fish were rarely seen
in these shallow habitats when temperatures increased.
It is interesting to note that while these individuals may
move to deeper, cooler habitats to avoid heat death,
the threat of predation is likely greater in these deep
pools where large predators (Micropterus spp.) were
observed on several occasions.
The presence of leopard darters in deeper water habitats has not been previously reported in the literature,
and our finding suggests that population estimates may
need to be reassessed. Recent analyses (Williams et al.
1999) of population viability for extant populations of
leopard darters have relied on estimates of availability
of suitable habitat as a basis for estimating total numbers within each of the tributaries of the Little River
system. In the model of Williams et al. (1999), suitable habitat was defined according to published studies
that noted leopard darters occupied pools less than 1 m
deep. Our observations of leopard darters using pools
up to 5 m deep during periods of warmwater temperature suggest that estimates of total numbers within
tributaries may need to take this habitat into account.
If survivorship in late summer depends on availability of adjacent deep pools that provide thermal refugia,
inclusion of only shallow pool habitats for population
estimates may substantially overestimate population
size. That is, some shallow habitats lacking access to
deeper pools nearby might not be suitable year-around
habitat, and including such habitat as ‘occupied’ would
inflate the total population estimate. Conversely, the use
of thermal refugia by leopard darters may cause population estimates based on late-summer censuses to be too
low. Reduction in population size that is often observed
in late summer in habitats that are adjacent to deep
pools (K. Collins unpublished data) may reflect migration into uncensused deep habitat rather than mortality.
Based on the results of field and laboratory studies and on data available in the literature, we made
the following recommendations to the USDA Forest
Service regarding construction/reconstruction of road
crossings in the Glover River and other streams and
rivers containing the leopard darter. Given that barriers appear to reduce movement of leopard darters, the
number of road crossings should be kept to a minimum
to facilitate local movement. However, if road crossings are necessary, they should be constructed to incorporate the widest possible box culverts, particularly if
construction of Denil fishways (which are known to be
used by other species of darters, Bunt et al. 2001), or
more elaborate fish passages are not an option. Most
importantly, habitat should be managed so that barriers
to local movement do not preclude access to thermal
refugia.
Acknowledgements
Funding for this project was provided by USDA
Forest Service Challenge Cost Share Agreement
#08-99-09-CCS-03. Leopard darters were collected
and housed under federal permit PRT-799158.
K. Collins, R. Standage and Sam Noble Oklahoma
Museum of Natural History provided valuable assistance with many aspects of this project. We thank
S. Ganick, A. Winkle, R. Durtsche, J. Hilliard,
M. Weddel and R. Marsh for assistance in the field.
References cited
Bunt, C.M., B.T. van Poorten & L. Wong. 2001. Denil fishway
utilization patterns and passage of several warmwater species
relative to seasonal, thermal and hydraulic dynamics. Ecol.
Freshwater Fish 10: 212–219.
Clapp, D.F., R.D. Clark, Jr. & J.S. Diana. 1990. Range, activity,
and habitat of large, free-ranging brown trout in a Michigan
stream. Trans. Amer. Fish. Soc. 119: 1022–1034.
Eley, R.L., J.C. Randolph & R.J. Miller. 1975. Current status
of the leopard darter, Percina pantherina. Southwest. Nat. 20:
343–354.
Freeman, M.C. 1995. Movements by two small fishes in a large
stream. Copeia 1995: 361–367.
Funk, J.L. 1955. Movement of stream fishes in Missouri. Trans.
Amer. Fish. Soc. 85: 39–57.
Gerking, S.D. 1959. The restricted movement of fish populations.
Biological Review 34: 221–242.
Gowan, C. & K.D. Fausch. 1996. Mobile brook trout in
two high-elevation Colorado streams: Re-evaluating the concept of restricted movement. Can. J. Fish. Aquat. Sci. 53:
1370–1381.
400
Harvey, B.C. 1991. Interactions among stream fishes: Predatorinduced habitat shift and larval survival. Oecologia 87: 29–36.
Hill, J.A. & G.D. Grossman. 1987. Home range estimators for
three North American stream fishes. Copeia 1987: 376–380.
James, P.W. 1996. Threatened fishes of the world: Percina
pantherina (Moore & Reeves, 1995) (Percidae). Env. Biol.
Fishes 43: 342
James, P.W. & O.E. Maughan. 1989. Spawning behavior and
habitat of the threatened leopard darter, Percina pantherina.
Southwest. Nat. 34: 298–301.
James, P.W., O.E. Maughan & A.V. Zale. 1991. Life history of the
leopard darter Percina pantherina in Glover River, Oklahoma.
Amer. Midl. Nat. 125: 173–179.
Jones, R.N., D.J. Orth & O.E. Maughan. 1984. Abundance and
preferred habitat of the leopard darter, Percina pantherina, in
Glover Creek, Oklahoma. Copeia 1984: 374–384.
Katapodis, C., A.J. Derksen & B.L. Christensen. 1991. Assessment of two Denil fishways for passage of freshwater species.
Amer. Fish. Soc. Symposium 10: 306–324.
Kaya, C.M., L.R. Kaeding & D.E. Burkhalter. 1977. Use of a
cold-water refuge by rainbow and brown trout in a geothermally
heated stream. Progr. Fish Cult. 39: 37–39.
Lonzarich, D.G., M.R. Lonzarich & M.L. Warren, Jr. 2000.
Effects of riffle length on the short-term movement of fishes
among stream pools. Can. J. Fish. Aquat. Sci. 57: 1508–1514.
Matthews, K.R. & N.H. Berg. 1997. Rainbow trout responses to
water temperature and dissolved oxygen stress in two southern
California stream pools. J. Fish Biol. 50: 50–67.
Matthews, W.J., B.C. Harvey & M.E. Power. 1994. Spatial
and temporal patterns in the fish assemblages of individual
pools in a midwestern stream (U.S.A.). Env. Biol. Fish. 39:
381–397.
Miller, R.J. & H.W. Robison. 1973. The Fishes of Oklahoma.
Oklahoma State University Press, Stillwater. 246 pp.
Peterson, J.T. & C.F. Rabeni. 1996. Natural thermal refugia for
temperate warmwater stream fishes. N. Am. J. Fish. Manag.
16: 738–746.
Power, M.E. 1984. Habitat quality and distribution of algaegrazing catfish in a Panamanian stream. J. Anim. Ecol. 53:
357–374.
Power, M.E., W.J. Matthews & A.J. Stewart. 1985. Grazing minnows, piscivorous bass and stream algae: Dynamics of a strong
interaction. Ecology 66: 1446–1456.
Schaefer, J.F. 1999. Movement by three stream-dwelling
cyprinids (Campostoma anomalum, Cyprinella venusta and
Notropis boops). Ph.D. Dissertation. University of Oklahoma,
Norman. 96 pp.
Schaefer, J.F. 2001. Riffles as barriers to interpool movement by
three cyprinids (Notropis boops, Campostoma anomalum and
Cyprinella venusta) Freshwater Biol. 46: 1–10.
Schlosser, I.J. 1995. Critical landscape attributes that influence
fish population dynamics in headwater streams. Hydrobiologia
303: 71–81.
Schwalme, K., W.C. Mackay & D. Linder. 1985. Suitability of
vertical slot and Denil fishways for passing north-temperate,
nonsalmonid fish. Can. J. Fish. Aquat. Sci. 42: 1815–1822.
Smith, R.K. & K.D. Fausch. 1997. Thermal tolerance and vegetation preference of Arkansas darter and johnny darter from
Colorado plains streams. Trans. Amer. Fish. Soc. 126: 676–686.
Smithson, E.B. & C.E. Johnston. 1999. Movement patterns of
stream fishes in a Ouachita Highlands stream: An examination
of the restricted movement paradigm. Trans. Amer. Fish. Soc.
128: 847–853.
Suren, A., G. Smart, R. Smith & S. Brown. 2000. Drag coefficients of stream bryophytes: Experimental determinations and
ecological significance. Freshwater Biol. 45: 309–317.
Toepfer, C.S., W.L. Fisher & J.A. Haubelt. 1999. Swimming performance of the threatened leopard darter in relation to road
culverts. Trans. Amer. Fish. Soc. 128: 155–161.
Tyus, H.M. 1990. Potamodromy and reproduction of Colorado
squawfish in the Green River basin, Colorado and Utah. Trans.
Amer. Fish. Soc. 119: 1035–1047.
U.S. Fish and Wildlife Service. 1978. Final threatened status
and critical habitat for the leopard darter. Federal Register 43:
3711–3716.
Warren, M.L., Jr. & M.G. Pardew. 1998. Road crossings as barriers to small-stream fish movement. Trans. Amer. Fish. Soc.
127: 637–644.
Williams, L.R., A.A. Echelle, C.S. Toepfer, M.G. Williams &
W.L. Fisher. 1999. Simulation modeling of population viability for the leopard darter (Percidae: Percina pantherina).
Southwest. Nat. 44: 470–477.
Zale, A.V., S.C. Leon, M. Lechener, O.E. Maughan, M.T. Ferguson,
S. O’Donnell, B. James & P.W. James. 1994. Distribution of the
threatened leopard darter, Percina pantherina (Osteichthyes:
Percidae). Southwest. Nat. 39: 11–20.